DSS DESIGN AND EVALUATION

1
DSS DESIGN AND EVALUATION

• Whats different about designing and developing a
  DSS vis-à-vis other kinds of information systems?
• How can we evaluate whether a DSS is providing
  any benefits?

2
SYSTEM LIFE CYCLE

• Planning
• Analysis
• Design
• Development
• Testing
• Evaluation
• Deployment
• Maintenance

3
DSS PROTOTYPE DEVELOPMENT PROCESS

• Start with well thought out statement of
  requirements
• Develop broad (but shallow) prototype with high
  productivity tools
• Interact with users to get feedback
• Enhance (deepen) prototype incorporating feedback
• Continue iterative process until acceptable
  production version emerges
• Continue adaptive enhancements during operational
  phase

4
THREE LEGS TO A SUCCESSFUL IMPLEMENTATION PROCESS

• Small, Talented Team
• Knowledge of implementation tools
• Ability to learn application domain
  characteristics
• Ability to contribute through entire system life
  cycle
• Ability to deal effectively with users

5
THREE LEGS (continued)

• Contemporary Software Development Tools
• Languages Java, C, VB
• Web-based XML, Java Script, ASP, ODBC
• Object-based UML, CORBA
• ADEs and CASE Tools PowerBuilder, Oracle
  Designer
• Component-oriented COTS software
• DSS Generators LDW, Analytica, Exsys

6
THREE LEGS (continued)

• Adaptive Methodology
• User-driven
• Informal coordination
• Complete access to users at all levels in the
  organization
• Authoritative support

7
DSS DESIGN CHARACTERISTICS ROMC

• Representations provide representations that
  support appropriate conceptualization of decision
  problem
• Operations support cognitive activities of
  formulation, analysis, and interpretation
• Memory Aids long-term and short-term
• Control Mechanisms control and direct the course
  of problem-solving

8
DSS COMPONENTS
Data Management

• Model
• Management

Knowledge Management
GUI
User
9
DATA MANAGEMENT

• Legacy System
• DBMS
• Relational
• SQL / QBE
• Dictionary / Repository
• Web-DBMS Development Tools
• Data Warehouse

10
MODEL MANAGEMENT

• Model Base
• Modeling Languages (AMPL GAMS SAS)
• Model Generators (e.g., Analytica LDW)
• Model Solvers
• Spreadsheets
• Database Interfaces
• Desirable Features
• What if
• Goal seeking
• Uncertainty (probability)
• Dynamic modeling

11
MODEL MANAGEMENT Modeling Life Cycle

• Requirements
• Model Formulation (modeling languages DSSG)
• Data Collection
• Model Solution (solvers e.g., simplex,
  regression)
• Sensitivity Analysis (what if, goal seeking)
• Model Calibration (is model realistic?)
• Model Validation (do users believe model
  results?)
• Deployment
• Maintenance and Enhancements

12
KNOWLEDGE MANAGEMENT

• Knowledge Acquisition
• Knowledge representation
• Knowledge Validation
• Desirable Features
• Inference engine (forward backward chaining)
• Explanation capability
• Truth maintenance

13
USER INTERFACES

• Input How does user manipulate DSS?
• Command Language
• Menus
• Q A
• Natural Language
• Voice
• GUI (e.g., GIS)

14
USER INTERFACES (continued)

• Output How are results of DSS presented to user?
• Graphics (e.g., charts)
• Spreadsheets
• Animation (e.g., simulations)
• Maps

15
MEASURES FOR DSS EVALUATION

• Productivity measures time and cost to make a
  decision (is decision making more efficient?)
• Process measures number of alternatives
  generated and examined (is decision making more
  effective?)
• Perception measures user friendliness
  confidence in results
• Product measures technical performance (e.g.,
  response time, reliability, etc)